| Due to the overload of information brought by a large amount of user-generated content(UGC),the popularity of the content in the community is currently determined by users through a voting mechanism,especially in social Q&A communities such as Quora.com in the United States and Zhihu in China.These Q&A sites have implemented effective information filtering by sorting answers based on the number of votes.The votes reflect the users' assessment of quality and influence their attention distribution.The fullness,richness and openness of the contents in e-commerce websites are far worse than UGC in the social Q&A community.It means that issues such as evaluating the quality of UGC in the social Q&A community and identifying antecedents of high reputation information will be more challenging than in e-commerce websites.The main research objective of this paper is to solve the following problems:(1)The source,connotation,formation mechanism,influence and analysis methods of UGC reputation characteristics in the social question-and-answer community;(2)Influencing factors of UGC's reputation characteristics in the social question-answering community and the relationship between these influencing factors and reputation characteristics;(3)Predict the reputation characteristics of UGC by using relevant observable variables of social question-answering community;(4)Explore the characteristics of UGC in different topics and reputation intervals;(5)Based on the relevant research on UGC reputation characteristics in the social question-and-answer community,this paper puts forward targeted management strategies and practical suggestions.Therefore,research on the UGC reputation characteristics of users in the social Q&A community is the basis for in-depth analysis of the formation mechanism of UGC reputation characteristics in the social Q&A community and identification of high reputation UGC information in advance.It helps to strengthen the position ofcommunities and governments in the evolution of public opinion and improve the governance level of the network information ecological environment.This research takes the user's UGC in the social Q&A community as the research object and starts with the user's reputation-giving behavior.Through theoretical analysis,empirical research and data mining methods,the following aspects are studied in depth: Social question-and-answer community user reputation gives behavior motivation,action mechanism,UGC reputation characteristics in the community,and analysis of the relationship between UGC information quality and source credibility.Moreover,the corresponding optimized strategy are proposed.This article is organized as follows:Firstly,I analyzed the motivations of users to give reputation in social Q&A community.Using grounded theory,combining Information Acceptance Model,explores the influential factors of UGC reputation acquisition in social Q&A community.Then combined with SOR theory,the formation mechanism of UGC reputation characteristics in social social Q&A community is studied.Then,the influence of UGC reputation characteristics is introduced.From the influence of UGC reputation characteristics on active users,"divers",social social Q&A community and Internet environment,the influence model of UGC reputation characteristics in social question-answering community is constructed.Finally,the UGC reputation characteristics of the social Q&A community are analyzed,the theoretical basis based on SOR theory is expounded,the analysis methods and analysis indexes of UGC reputation characteristics of the social Q&A community are explained,and the UGC reputation characteristic correlation theory model of users in the social question-answering community is constructed.Secondly,based on the information acceptance model,this paper constructs a model of influencing factors of UGC reputation characteristics in social Q&A community.It proposes that UGC text characteristics,non-text characteristics and source credibility in social Q&A community are the influencing factors of UGC reputation characteristics and are regulated by information involvement.Then,it takes190,990 pieces of answer data on the three topics of health,fashion and entertainmentin the community as an example for empirical research.The results show that:(1)for all topics,the better the quality of the text,the higher the reputation of the answer.Text features,non-text features and source credibility all have positive effects on the size of reputation features,but this effect is very small for the answer under entertainment topics;(2)In most cases(including the general situation,health and fashion topics),the higher the credibility of the source,the less likely it is to be zero for the response reputation characteristics,but entertainment topics are an exception;(3)There is no moderating effect of various factors of topic involvement on reputation characteristics;(4)There is a phenomenon of "structural zero" in UGC reputation features of social Q&A communities.There is a certain difference in the influence of source credibility on the "structural zero" and discrete distribution parts of reputation features.Thirdly,through literature research,the index attributes that affect the quality of answer information and the credibility of information sources are constructed from the relevant attributes of answer information in social Q&A communities.The information quality includes two aspects: text features and non-text features.Text features include four indicators of answer length,emotional polarity,keyword density and question-and-answer length ratio.Non-text features include six indicators of timeliness of answer,timeliness of feedback,number of comments,interactivity of respondents,number of quoted external links and number of pictures.The indicators of source credibility include the number of respondents' praises,the number of respondents' collections,the best number of respondents' answers,the personal authentication of respondents,the official authentication of respondents,the number of respondents' fans,the number of respondents' concerns,the number of respondents' answers,the number of respondents' articles,and the total number of respondents' professionalism.Then,the method of obtaining and quantifying the relevant indexes of three variables,namely,information quality text feature,non-text feature and source credibility,is proposed.The UGC reputation prediction model of social question-and-answer community is constructed.The influence coefficients of the three variables of information quality text features,non-text features and sourcecredibility on the information reputation features are taken as weights,and UGC reputation features are predicted through information quality and source credibility.The prediction of this model includes two aspects: one is to predict the reputation characteristics that may be obtained by the existing reputation answers;The second is to predict whether the zero reputation state of answering the zero reputation answer will continue.Fourthly,the answer data from Zhihu are selected for example verification.After verifying the prediction effect of the model,this chapter also explores the distribution of UGC information quality and source credibility in different reputation intervals,and analyzes UGC with abnormal reputation characteristics in the sample.(1)The main influencing dimension of text features is the ratio of text length to question-and-answer length;For non-text feature dimensions,the influence weight of feedback timeliness is relatively low.The reward obtained by the respondents by publishing UGC far affects the credibility of the respondents more than the number of UGC published by the respondents,and the number of articles written by the respondents also has a greater impact than the number of answers by the respondents.(2)There are great differences in the attributes of vocabulary in different topics.The proportion and relative position of emotional words in high-frequency characteristic words answered in high reputation and extremely high reputation intervals in health topics are smaller,while subjective words in fashion and entertainment topics are closer to the center.(3)In the social Q&A community,UGC's source credibility determines the lower limit of the response's reputation characteristics,but cannot determine its upper limit.Reasonable surface language features and non-text features are the necessary conditions to obtain a high reputation for answering questions.(4)The authentication of the respondents can only improve the reputation characteristics of the current answer information,but cannot improve the overall reputation characteristics of all the answers within the topic.Fifth,from the three levels of users,social Q&A community management and government network information governance,the corresponding management strategies are put forward.At the user level,while guiding users to publish orcooperate with others to produce UGC with higher quality and more influence,it advocates users' normative reputation granting behavior and puts forward specific suggestions.At the level of social Q&A community management,feasible strategies are put forward in the practical measures to encourage high-quality UGC and high-quality users,and reference is provided for optimizing the ranking mechanism of UGC in the community.At the level of government information governance,we will help the government to "stir up the turmoil" so that high-quality opinion leaders and UGC can play their role in Internet information governance earlier,more widely and more deeply,and rumors and junk information that may have certain influence can be properly dealt with before they cause serious consequences.At the theoretical level,this study enriches and improves the theoretical system of UCG research,expands and deepens the theoretical framework of UCG information characteristics,and provides scientific methods and theoretical basis for the ranking optimization of UCG in social Q&A communities.On the practical level,by analyzing the relevant rules of the UGC reputation characteristics of the social Q&A community,this research provides a systematic practical method for screening and identifying high-reputation user-generated content.This helps to strengthen the emergence of user-contributed knowledge in socialized Q& A community,and is also conducive to promoting the healthy development of the socialized Q&A community. |
PDF Full Text Request